xAI’s Grok Vision lets iPhone users ask questions about what they point their camera at. New voice features also support multiple languages and real-time search. xAI has launched a new feature called Grok Vision for its chatbot, Grok. This update allows iPhone users to use their phone’s camera to interact with the world in real time. Users can point their camera at an object—like a product, sign, or document—and ask Grok questions about what they see. The chatbot responds using its visual understanding. This brings Grok in line with similar…

Google’s A2A protocol and HyperCycle’s platform aim to standardise AI agent communication and scale secure, cross-platform collaboration. AI agents are becoming more capable, handling complex tasks like managing supply chains. But as organizations use agents from different vendors, these systems often can’t communicate or work together. This creates silos and conflicting outcomes. Integrating agents usually needs extra software, making systems more complex and prone to errors. To fix this, Google introduced the Agent2Agent (A2A) protocol at Cloud Next 2025. A2A is open-source and lets AI agents from different platforms and…

April 2025 saw an explosion of activity in the AI industry. Major companies released cutting-edge AI new technologies.Below, we highlight the most significant launches and announcements, drawing insights from top news and official sources. This article was sourced from AINEWS. For more details, please read: Google launches A2A as HyperCycle advances AI agent interoperability Google’s A2A Protocol and HyperCycle Boost AI Agent Collaboration AI agents are becoming more capable, handling complex tasks like managing supply chains. But as organizations use agents from different vendors, these systems often can’t communicate or…

OpenAI launches o3 and o4-mini AI models, boosting reasoning, coding, and image analysis. Now available to ChatGPT Pro, Plus, and Team users. OpenAI has released two new AI reasoning models: o3 and o4-mini. These models are designed to pause and think before answering, making their replies more thoughtful and accurate. o3 is OpenAI’s most advanced reasoning model yet. It outperforms older models in math, coding, science, and image understanding. o4-mini is a lighter model that offers a good balance between speed, cost, and performance. There’s also a version called o4-mini-high,…

S2 by Simular blends top AI models for smarter, multi-step task handling on computers and phones. It’s more accurate, flexible, and learns from feedback. S2 is a new AI agent made by a startup called Simular. This agent is special because it uses more than one AI model. It picks the best model depending on the task. This makes it more flexible and smarter than older agents. S2 mixes powerful general AI models like GPT-4o with smaller, focused models. The general model handles planning. The smaller ones help with things…